The Girl Who Played With Statistics: A Cautionary Tale

On a recent post, a commenter that goes by the name Clarence asked me what I thought of a certain study. It’s called The Harmful Effects of Early Sexual Activity And Multiple Sexual Partners Among Women: A Book of Charts. The report contains 18 charts that demonstrate that, you guessed it, early sexual activity and multiple sexual partners among women has harmful effects!

The study was released by The Heritage Foundation, an extreme right-wing lobby group. The four authors of the study (only one of whom has any academic credentials) all work for The Heritage Foundation. According to them, the data for these charts comes from the CDC’s 1995 National Survey of Family Growth. The CDC being a credible and generally unbiased institution, this seemed a bit fishy.

Clarence said that he couldn’t find any “definitive debunking” of The Heritage Foundation’s terrifying conclusions and that the only way to get to the bottom of this was to study the methodology of the CDC’s 1995 National Survey, and “make sure the Heritage people aren’t trying to pull a fast one by misrepresenting something.” When I suggested that “pulling a fast one by misrepresenting something” was exactly what The Heritage Foundation was doing, Clarence wrote, “I’m not a fan of The Heritage Foundation, but I can’t dismiss a study like this based solely on who put it out.”

But sometimes, folks, it really is in your best interest to dismiss a study based solely on who put it out, especially if it’s The Heritage Foundation. Seriously. For I am about to share a cautionary tale of a young woman who did NOT dismiss this study. Nay, she tried to get to the bottom of it, completely unaware of the risk to her own health and sanity she was taking. This is the story of my good friend, who wished to remain anonymous, in her own words.

A fever- fueled hysterical breakdown

I like stats. Reading this comments section spurred one of those crazy stats urges.  Pre-martial sex and the stability of marital partnerships! Social theory with numbers!  I generally like and trust the CDC when it comes to numbers-making.  In my mind they are a shiny quarter-fed number machine.

Looking at the first chart, it seemed a little strange to me that there would only be marriages recorded for women over 30 since the CDC’s data covers ages 15-44.  I also knew that the CDC keeps record of religious belief and other fun stuff.  Maybe those people who are more religious are a large portion of the ones who have fewer partners and understandably also prioritize avoiding divorce? Or maybe cohabitation has replaced marriage for lots of people?  So many fun things that this graph might be hiding in its finer points.

…Four hours later of reading through everything I never knew I never knew about marriage in 1995, I realized I had to stop. I had found the CDC pages that were probably referenced within the first thirty minutes and nothing else was coming forward.

They are here, here, and here.

If there is someone reading this who knows what I’m missing please read them and tell me what it is.

I tried to wrap my brain around how the Heritage Foundation got to the stats they did. It was terrible. I thought I was playing a fun game of find the numbers and discover how they came to be, or how they might have been misrepresented.

There was no game.
No rules.

Either the chart was made up and has nothing to do with the original numbers or I’ve lost my ability to read coherently.  I couldn’t even find any stats that cover more than 10 partners. Where are the 11-20+ partners? Where are the 80% of stable first partner relationships?

These pages, Leah, I think they are where the answer is. I just want someone to tell me for the love of god where the numbers are. I think I would almost be relieved. I also have a fever from flu, which might be driving my hysteria. But that doesn’t make me dumb.

Please find the numbers for me.  I’m trapped in one of those schizophrenic mathematician movies in the final scenes where the universe is created by their own mind and there are no numbers really there at all.  I would also be super-sad and way too naively surprised if the graph was actually not drawn from the stats on the CDC’s report at all, but after searching and searching, I’m afraid I must conclude that they made it up.

This, my friends, is why it’s so hard to argue against the anti-sex, anti-feminist lobby’s use of “data” and “studies” and “surveys” and “statistics.” It’s because, more often than not, their data is absolute bullshit. Groups like The Heritage Foundation and people like Susan Walsh are pushing extreme political agendas based on an outdated model of sexist morality. They believe they are “battling” to preserve the future of society as we know it, and since all is fair in love and war, they have no problem bending, twisting, appropriating, or even fabricating data that supports their cause. It’s hard to fight bullshit with logic, and as my poor friend’s descent into madness demonstrates, the psychological effects of attempting to do so can be severe.

So, folks, next time someone asks you to “definitively debunk” a “Book of Charts” disseminated by The Heritage Foundation, go ahead and dismiss it solely because it was put out by The Heritage Foundation. Save yourself. It’s just not worth it.

As for my dear friend, I wish her a speedy recovery. Maybe Delusions of Gender, a promising new book in which a cognitive neuroscientist uses good science to identify and eviscerate those who do not, could provide some much-needed therapy for her poor, stats-loving soul.

5 thoughts on “The Girl Who Played With Statistics: A Cautionary Tale

  1. Hi

    The tables for the most wanted numbers (not for the entire “study”) are: Series 23, no 19, page 41; but only around 36% of all women with only one partner in their lifetime are still married. And the 80 % stable first partner relationships comes from “About 3 out of 5 women (61 percent) were ‘‘going steady’’ or ‘‘going together’’ with the man they had intercourse with the first time, and about 1 in 5 were engaged or married to him.” (page 5 and in numbers, but more confusingly: page 35 in the same report), although to speak of stable first partners is wrong, as this question can only be answered after the death of the women.

    regards,
    Simon

  2. These studies always sound sort of like the one that Heather Corinna (sort of) debunked about two weeks ago on Scarleteen. Either the numbers are completely made up (and Echidne is pretty good at clearing up those statistical issues) or the conclusions the organizations come to are WILDLY out of bounds with what the data actually shows.

  3. Pingback: Now, here’s a sex study that makes sense | Not a Dirty Word

  4. I too have tried to rectify the findings in the Heritage “study” with the CDC data. The best that I could come up with is that the Heritage people took narrow slices of data to create statistically significant findings. The weird part is, I was surprised they had to even do that. I mean, I believe there probably is a correlation between low numbers of pre-marital partners and marriage success. But a simple correlation is far from the oft-repeated “rule” in the manosphere that “sluts lose their ability to pair-bond.”

    I should carry a billy club to whack anybody who starts any argument with “studies show…” and an even bigger bat for people who reference “studies” that aren’t even studies.

  5. I’m disturbed too and want to get to the bottom of this, since I want to question the validity of the findings. Especially chart 15 titled: “Women Who Have More Non-Marital Sexual Partners Are Less Likely to Have Stable Marriages”

    I suspect the Heritage Foundation used the dataset of survey responses, not the summary data tables linked in the article above. Since, none of the tables in the 3 links posted have:

    1)Average # of Partners
    2)#/% of women in 5+ year long current marriages
    ..on the same table.

    In order to create that table, an analyst MUST go to the source data. You can’t just combine summary tables like proposed by your friend.

    For example if I said average apple trees grows 30 fruit and 50% of all Apples were grown in USA, that’s useless. There’s no way to connect the two. You can’t say average US apple tree grows 15 fruit. You can’t say anything new with it. There’s just no way of combining those 2 pieces of information. Likewise if I said women had 2 partners by 30, and 50% of age 30 women were in 5+year marriages, you can’t say anything. You can’t say average women in 5 year marriages had 1 partner.

    Trying to use the tables above would be attempting that.

    To get to the answer, you need to create a new table using the survey responses themselves. Because there, each individual participant has said: “I have had X many partners” and “My current marriage is X years old”

    I know this because here is the survey: http://www.cdc.gov/nchs/data/nsfg/nsfgqx95.pdf
    Here is the dataset of responses: ftp://ftp.cdc.gov/pub/Health_Statistics/NCHS/Datasets/NSFG/1995FemRespData.dat

    In order to double check the Heritage Foundation’s #’s you’d need to open that dataset in SAS and rerun those tables. Unfortunately I don’t know how to use SAS. If someone sees this who can, that’d be much appreciated.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s